B1718
Title: PLreg: an R package for modeling bounded continuous data
Authors: Francisco F Queiroz - University of Sao Paulo (Brazil) [presenting]
Silvia Ferrari - University of Sao Paulo (Brazil)
Abstract: The power logit class of distributions is useful for modelling continuous data on the unit interval, such as fractions and proportions. It is very flexible and the parameters represent the median, dispersion and skewness of the distribution. The power logit regression models are based on the power logit class. The dependent variable is assumed to have a distribution in the power logit class with its median and dispersion linked to regressors through linear predictors with unknown coefficients. The power logit class of distributions and the associated regression models are implemented in the R package PLreg. The methods and algorithms implemented in the package are described and illustrated, including parameter estimation, diagnostic tools associated with the fitted model as well as density, cumulative distribution, quantile, and random number generating functions of the power logit distributions. Additional illustrations are presented to show the ability of the PLreg package to fit generalized Johnson SB, log-log, and inflated power logit regression models.
